Identification of H-specific determinants in flagellin of four Escherichia coli strains

The H serogroup of Escherichia coli is determined by the flagellar antigen, flagellin. Sequence analysis of the flagellin gene, fliC, reveals a central variable region and the highly conserved N- and C-termini. This variable region has been shown to encode both H-specific and cross-reactive epitopes. Using polyclonal antibodies, we mapped the linear H-specific determinants in flagellin from four E. coli serotypes O157:H10, 0138:H14, O157:H42 and O157:H43. The specificity of all potential fragments was verified with 52 ECRC (Escherichia coli Reference Center) H-specific antisera. Our results indicated that: (a) a specific determinant of H10 flagellin (1263 bp long) maps to the region covering amino acid residues 305-331; (b) a specific determinant of H14 flagellin (1653 bp long) maps to the region covering amino acid residues 430-461; (c) a specific determinant of H42 flagellin (1281 bp long) maps to a region covering amino acid residues 171-201; and (d) a specific determinant of H43 flagellin (1506 bp long) maps to a region covering amino acid residues 200-260.     

Involvement of tyrosine residues in the protomer-protomer interaction of Proteus mirabilis flagella as studied by spectroscopic methods, chemical modification and aggregation experiments.

Using spectrophotometrical titration, chemical modification, and ultraviolet difference spectral methods, the existence of at least two distinct tyrosine groups in the isolated flagellin of Proteus mirabilis flagella has been established. Three of the five flagellin tyrosines are buried in the protein matrix, whereas the other two seem to lie on the protein surface accessible to perturbants. Also about two tyrosine residues, presumably the latter ones exposed to the environment, can be nitrated with tetranitromethane in the monomeric flagellin with a concomitant loss of the polymerization ability after about one tyrosine per mol flagellin has been nitrated. Nitrated flagellin, homogeneous with respect to molecular weight, degree of nitration and isoelectric point, could be isolated and characterized. On the other hand, it could be shown that in the polymeric flagellum the phenolic groups of all five tyrosine residues are inaccessible to perturbing and modifying reagents. It seems, therefore, that the integrity of the phenolic groups is necessary for the proper folding and aggregation of the flagellin subunits to form the stable helical flagella.     

Amino Acid Substitutions and Intragenic Duplications of Bacillus sp. PS3 Flagellin Cause Complementation of the Bacillus subtilis Flagellin Deletion Mutant

Bacillus sp. PS3 produces a glycosylated flagellin. In this study, a number of the glycosylated residues of the flagellin protein were found to be located in the central variable region of this protein. We also report that the motility defect of the Bacillus subtilis flagellin mutant was complemented by Bacillus sp. PS3 flagellin variants without glycosylation, which contained amino acid substitutions and intragenic duplications in the variable region of flagellin.     

Identification and mutational analysis of Arabidopsis FLS2 leucine-rich repeat domain residues that contribute to flagellin perception

Mutational, phylogenetic, and structural modeling approaches were combined to develop a general method to study leucine-rich repeat (LRR) domains and were used to identify residues within the Arabidopsis thaliana FLAGELLIN-SENSING2 (FLS2) LRR that contribute to flagellin perception. FLS2 is a transmembrane receptor kinase that binds bacterial flagellin or a flagellin-based flg22 peptide through a presumed physical interaction within the FLS2 extracellular domain. Double-Ala scanning mutagenesis of solvent-exposed beta-strand/beta-turn residues across the FLS2 LRR domain identified LRRs 9 to 15 as contributors to flagellin responsiveness. FLS2 LRR-encoding domains from 15 Arabidopsis ecotypes and 20 diverse Brassicaceae accessions were isolated and sequenced. FLS2 is highly conserved across most Arabidopsis ecotypes, whereas more diversified functional FLS2 homologs were found in many but not all Brassicaceae accessions. flg22 responsiveness was correlated with conserved LRR regions using Conserved Functional Group software to analyze structural models of the LRR for diverse FLS2 proteins. This identified conserved spatial clusters of residues across the beta-strand/beta-turn residues of LRRs 12 to 14, the same area identified by the Ala scan, as well as other conserved sites. Site-directed randomizing mutagenesis of solvent-exposed beta-strand/beta-turn residues across LRRs 9 to 15 identified mutations that disrupt flg22 binding and showed that flagellin perception is dependent on a limited number of tightly constrained residues of LRRs 9 to 15 that make quantitative contributions to the overall phenotypic response.     

Cleavage of bacterial flagellin with cyanogen bromide. Chemical and physical properties of the protein fragments

1. Flagellin, isolated from the flagella of Salmonella adelaide, was shown by various criteria to be a pure protein. It had a molecular weight of about 40000 and contained three methionine, six tyrosine, 11 arginine and 25 lysine residues/mol., of which 11 of the lysine residues were present as in-N-methyl-lysine. 2. After treatment of flagellin with cyanogen bromide in formic acid, four main fragments (A, B, C and D) were obtained, with as many as six minor components that represented partial degradation products. The major fragments were estimated by amino acid analysis to have molecular weights of about 18000 for fragment A, 12000 for fragment B, 5500 for fragment C and 4500 for fragment D. Fragments A, B and D, but not fragment C, were recovered pure by gel chromatography as monitored by polyacrylamide-gel electrophoresis. 3. A complex between fragments C and D was also isolated (mol.wt. 10000) after limited oxidation of flagellin by chloramine-t before digestion by cyanogen bromide. After oxidation essentially only two fragments were released from flagellin by cyanogen bromide: the ;C,D' complex and a presumed ;AB' fragment. 4. The sum of the amino acid analyses of fragments A and B and the ;C,D' complex gave residue values that agreed well with the amino acid composition of native flagellin. 5. Fragments A and D contained tyrosine, and ten of the 11 in-N-methyl-lysine residues of the molecule were in fragment A. Reaction with [(125)I]iodide at small extents of substitution showed that, in flagellin, the tyrosine residue of fragment D was more readily substituted than those of fragment A. By contrast, in polymerized flagellin, the tyrosine residues of fragment A were more readily substituted. 6. Treatment of flagellin with carboxypeptidases A and B revealed the C-terminal sequence -Leu-Leu-Leu-Arg. Arginine and leucine were released by carboxypeptidase from the ;C,D' complex but not from fragment D, indicating that fragment C was C-terminal. 7. On the basis of the results from amino acid analysis, carboxypeptidase digestion, N-terminal analysis, iodination studies and polyacrylamide-gel electrophoresis, the sequence of fragments in flagellin was considered to be B-A-D-C; in the polymer, fragment A was exposed. It is suggested that methylation of the lysine residues occurred in the organism after flagellin had polymerized.     

Identification of a sequence in human toll-like receptor 5 required for the binding of Gram-negative flagellin

Flagellins from Gram-negative bacteria activate inflammatory cells by a toll-like receptor 5 (TLR5)-dependent signaling pathway. We have examined the interaction between flagellin and TLR5 using an in vitro binding assay. Purified recombinant His-tagged flagellin from Salmonella enteritidis bound to TLR5 in detergent lysates from COS-1 cells transiently transfected with a human TLR5 expression plasmid. Flagellins from Salmonella typhimurium and Escherichia coli also bound to TLR5. The specificity of this interaction was demonstrated by its concentration dependence and lack of TLR5 binding to a biologically inactive form of flagellin or to a His-tagged non-flagellar protein. Flagellin bound to the extracellular domain of TLR5 expressed on the surface of COS-1 cells and to a soluble, monomeric form of the extracellular domain (amino acids 1-636). Although a TLR5 extracellular domain containing amino acids 1-407 retained flagellin binding activity, binding was not evident with a TLR5 peptide encoding residues 1-386. Conversely, a peptide containing amino acid residues 386-636 retained flagellin binding. Thus it is likely that amino acids 386-407 is a binding site for flagellin. This sequence contains a putative leucine-rich repeat. These results support the conclusion that flagellin signaling via TLR5 involves a direct interaction between flagellin and a leucine-rich region in TLR5. We also show that the NH2-terminal 358 amino acids of TLR5 play an important role in its signaling activity. Our results provide, for the first time, a molecular basis for the agonist specificity of a TLR.     

Thermal perturbation difference spectroscopy of Salmonella flagellin

The appropriateness of a two-state model for the previously reported thermal transition of $\text{Salmonella}$ flagellin and the reversibility of this process recently has been questioned by others. Spectrophotometric evidence is presented here that reveals two separate thermally-induced structural transitions in flagellin which may resolve apparent controversy. A low temperature transition (I) appears between 28 and 35°C with a midtransition temperature of 30.7°C. With increasing temperature in this transition region a progressive shift to the red of the absorbance band at 284 nm occurs. The latter, probably due to an infolding of tyrosyl residues, is paralleled by a decrease in the rate of polymerization of flagellin. The temperature profile for the spectral behavior of flagellin in transition I fulfills criteria for a two-state process and is fully reversible. A second transition, also reversible, is observed between 40 and 60°C with a midtransition temperature near 50°C. Transition II, observed as a blue spectral shift of the absorbance band at 277 nm, is better described as a multistate process. Tyrosyl residues appear to maintain the conformational integrity of the polymerizable state of flagellin.     

Evidence for posttranslational modification and gene duplication of Campylobacter flagellin.

A gene encoding a flagellin protein of Campylobacter coli VC167 has been cloned and sequenced. The gene was identified in a pBR322 library by hybridization to a synthetic oligonucleotide probe corresponding to amino acids 4 to 9 of the N-terminal sequence obtained by direct chemical analysis (S. M. Logan, L. A. Harris, and T. J. Trust, J. Bacteriol. 169:5072-5077, 1987). The DNA was sequenced and shown to contain an open reading frame encoding a protein with a molecular weight of 58,945 and a length of 572 amino acids. The deduced amino acid sequence was identical to the published N-terminal amino acid sequence of VC167 flagellin and to four internal regions whose partial sequences were obtained by direct chemical analysis of two tryptic and two cyanogen bromide peptides of VC167 flagellin. The C. coli flagellin protein contains posttranslationally modified serine residues, most of which occur within a region containing two 9-amino-acid repeating peptides separated by 34 unique amino acids. Comparisons with the sequences of flagellins from other bacterial species revealed conserved residues at the amino- and carboxy-terminal regions. Hybridization data suggest the presence of a second flagellin copy located adjacent to the first on the VC167 chromosome.     

Structural organization of flagellin

The terminal regions of flagellin from Salmonella typhimurium have been reported to be disordered in solution, whereas the central part of the molecule contains protease-resistant, compact structural units. Here, conformational properties of flagellin and its proteolytic fragments were investigated and compared to characterize the domain organization and secondary structure of flagellin. Deconvolution analysis of the calorimetric melting profiles of flagellin and its fragments suggests that flagellin is composed of three co-operative units or domains. The central part of the molecule, residues 179 to 418, consists of two domains (G1 and G2), whereas the third domain (G3) is discontinuous, constructed from segments 67 to 178 and 419 to 448. Secondary structure prediction and analysis of far-ultraviolet circular dichroic spectra have revealed that G1 and G2 consist predominantly of beta-structure with a little alpha-helical content. G3 contains almost equal amounts of alpha and beta-structure, while in the terminal parts of flagellin the ordered secondary structure seems to be entirely alpha-helical.     

Identification of the carbohydrate moieties and glycosylation motifs in Campylobacter jejuni flagellin.

Flagellins from three strains of Campylobacter jejuni and one strain of Campylobacter coli were shown to be extensively modified by glycosyl residues, imparting an approximate 6000-Da shift from the molecular mass of the protein predicted from the DNA sequence. Tryptic peptides from C. jejuni 81-176 flagellin were subjected to capillary liquid chromatography-electrospray mass spectrometry with a high/low orifice stepping to identify peptide segments of aberrant masses together with their corresponding glycosyl appendages. These modified peptides were further characterized by tandem mass spectrometry and preparative high performance liquid chromatography followed by nano-NMR spectroscopy to identify the nature and precise site of glycosylation. These analyses have shown that there are 19 modified Ser/Thr residues in C. jejuni 81-176 flagellin. The predominant modification found on C. jejuni flagellin was O-linked 5,7-diacetamido-3,5,7,9-tetradeoxy-l-glycero-l-manno-nonulosonic acid (pseudaminic acid, Pse5Ac7Ac) with additional heterogeneity conferred by substitution of the acetamido groups with acetamidino and hydroxyproprionyl groups. In C. jejuni 81-176, the gene Cj1316c, encoding a protein of unknown function, was shown to be involved in the biosynthesis and/or the addition of the acetamidino group on Pse5Ac7Ac. Glycosylation is not random, since 19 of the total 107 Ser/Thr residues are modified, and all but one of these are restricted to the central, surface-exposed domain of flagellin when folded in the filament. The mechanism of attachment appears unrelated to a consensus peptide sequence but is rather based on surface accessibility of Ser/Thr residues in the folded protein.     

Flagellin glycans from two pathovars of Pseudomonas syringae contain rhamnose in D and L configurations in different ratios and modified 4-amino-4,6-dideoxyglucose

Flagellins from Pseudomonas syringae pv. glycinea race 4 and Pseudomonas syringae pv. tabaci 6605 have been found to be glycosylated. Glycosylation of flagellin is essential for bacterial virulence and is also involved in the determination of host specificity. Flagellin glycans from both pathovars were characterized, and common sites of glycosylation were identified on six serine residues (positions 143, 164, 176, 183, 193, and 201). The structure of the glycan at serine 201 (S201) of flagellin from each pathovar was determined by sugar composition analysis, mass spectrometry, and (1)H and (13)C nuclear magnetic resonance spectroscopy. These analyses showed that the S201 glycans from both pathovars were composed of a common unique trisaccharide consisting of two rhamnosyl (Rha) residues and one modified 4-amino-4,6-dideoxyglucosyl (Qui4N) residue, beta-D-Quip4N(3-hydroxy-1-oxobutyl)2Me-(1-->3)-alpha-L-Rhap-(1-->2)-alpha-L-Rhap. Furthermore, mass analysis suggests that the glycans on each of the six serine residues are composed of similar trisaccharide units. Determination of the enantiomeric ratio of Rha from the flagellin proteins showed that flagellin from P. syringae pv. tabaci 6605 consisted solely of L-Rha, whereas P. syringae pv. glycinea race 4 flagellin contained both L-Rha and D-Rha at a molar ratio of about 4:1. Taking these findings together with those from our previous study, we conclude that these flagellin glycan structures may be important for the virulence and host specificity of P. syringae.     

Role of the disordered terminal regions of flagellin in filament formation and stability

Terminal regions of flagellin from Salmonella typhimurium, residues 1 to 65 and 451 to 494, have no ordered tertiary structure in solution, which makes them very susceptible to proteolytic degradation. Flagellin was subjected to mild controlled proteolytic treatment with highly specific proteases to remove terminal segments from the disordered regions. It is demonstrated here that various fragments can be readily prepared that differ from each other in 1 x 10(3) to 2 x 10(3) Mr segments in their NH2- or COOH-terminal regions. Terminally deleted fragments of flagellin were used to clarify the role of the disordered regions in the self-assembly of flagellin. The polymerization ability of the fragments was tested by inducing filament formation with ammonium sulfate. We found that fragments of flagellin containing large terminal deletions could form straight filaments, although the stability of these filaments required high salt concentrations. Even a fragment lacking the whole mobile COOH-terminal part of flagellin and 36 residues from the NH2-terminal region could form long filaments. The fragments could be also polymerized onto native flagellar seeds, suggesting that the subunit packing of the filaments of fragments is similar to that of the native ones. The fragments could also copolymerize with native flagellin, resulting in various helical forms. Filaments of fragments were found to be straight at both pH 4.0 and pH 12.5, indicating that they might have lost their polymorphic ability. Our results show that the major part of the disordered terminal regions of flagellin is not essential for polymerization, but it does play an important role in stabilization of the filaments and in influencing their polymorphic conformation.     

Structural and antigenic characteristics of Campylobacter coli FlaA flagellin.

The polar flagellar filament of Campylobacter coli VC167 is composed of two highly related (98%) flagellin subunit proteins, FlaA and FlaB, whose antigenic specificities result from posttranslational modification. FlaA is the predominant flagellin species, and mutants expressing only FlaA form a full-length flagellar filament. Although the deduced M(r) of type 2 (T2) FlaA is 58,884 and the apparent M(r) by sodium dodecyl sulfate-polyacrylamide gel electrophoresis is 59,500, the solution weight-average M(r) by sedimentation analysis was 63,000. Circular dichroism studies in the presence or absence of 0.1% sodium dodecyl sulfate or 50% trifluorethanol showed that the secondary structure of T2 FlaA flagellin was altered, with alpha-helix structure being increased to 25% in the nonpolar environment. The molecule also contained 35 to 48% beta-sheet and 11 to 29% beta-turn structure. Mimeotope analysis of octapeptides representing the sequence of FlaA together with immunoelectron microscopy and enzyme-linked immunosorbent assay with a panel of antisera indicated that many residues in presumed linear epitopes were inaccessible or nonepitopic in the assembled filament, with the majority being in the N-terminal 337 residues of the 572-residue flagellin. Residues at the carboxy-terminal end of the T2 FlaA subunit also become inaccessible upon assembly. Digestion with trypsin, chymotrypsin, and endoproteinase Glu-C revealed a protease-resistant domain with an approximate M(r) of 18,700 between residues 193 and 375. Digestion with endoproteinase Arg-C and endoproteinase Lys-C allowed the mapping of a segment of surface-exposed FlaA sequence which contributes serospecificity to the VC167 T2 flagellar filament at residues between 421 and 480.     

The amino acid composition of Rhizobium meliloti flagellin: Occurrence of a single tryptophan residue

Flagella were isolated from cultures of $\text{Rhizobium meliloti}$ strain RM2011 and purified to homogeneity as determined by the presence of a single band of the monomeric flagellin subunit following sodium dodecyl sulfate polyacrylamide gel electrophoresis. Classical hydrochloric acid hydrolysis showed that the amino acid composition of RM2011 flagellin was generally similar to that of other bacterial flagellins. Hydrolysis with methanesulfonic acid indicated that, unlike other flagellins, RM2011 flagellin contained a single tryptophan residue per molecule. Analysis of the products obtained by the reaction of the flagellin with N-bromosuccinimide under conditions specific for cleavage at tryptophan residues confirmed the presence of a single tryptophan residue.     

Production of a monoclonal antibody specific to Escherichia coli H33 flagellin by using predetermined polypeptides

The Escherichia coli H serogroup is determined by flagellin, which has both H-type-specific and cross-reactive epitopes. The cross-reactive epitopes are responsible for the cross-reaction found in agglutination. To identify the specific epitope in H33 flagellin, the H33 flagellin gene was sequenced and the encoded central variable region (CVR) was determined. Four overlapping fragments of the CVR were prepared and their specificity was verified using different H-type antisera. Short fragments carrying potential H-type-specific determinants were selected, and monoclonal antibodies (MAbs) against these fragments were prepared. A murine MAb of subtype IgG1 showing specificity to H33 flagellin was produced. The epitope of the MAb was mapped to amino acid residues 250-260.     

A deletion variant study of the functional role of the Salmonella flagellin hypervariable domain region in motility

The eubacterial flagellum is a complex structure with an elongated extracellular filament that is composed primarily of many subunits of a flagellin protein. The highly conserved N and C termini of flagellin are important in its export and self-assembly, whereas the middle sequence region varies greatly in size and composition in different species and is known to be deletion-tolerant. In Salmonella typhimurium phase 1 flagellin, this "hypervariable" region encodes two solvent-exposed domains, D2 and D3, that form a knob-like feature on flagella fibers. The functional role of this structural feature in motility remains unclear. We investigated the structural and physiological role of the hypervariable region in flagella assembly, stability and cellular motility. A library of random internal deletion variants of S. typhimurium flagellin was constructed and screened for functional variants using a swarming agar motility assay. The relative cellular motility and propulsive force of ten representative variants were determined in semi-solid and liquid medium using colony swarming motility assays, video microscopy and optical trapping of single cells. All ten variants exhibited diminished motility, with varying extents of motility observed for internal deletions less than 75 residues and nearly complete loss of motility for deletions greater than 100 residues. The mechanical stability of the variant flagella fibers also decreased with increasing size of deletion. Comparison of the variant sequences with the wild-type sequence and structure indicated that all deletions involved loss of hydrophobic core residues, and removal of both partial and complete segments of secondary structure in the D2 and D3 domains. Homology modeling predicted disruptions of secondary structures in each variant. The hypervariable region D2 and D3 domains appear to stabilize the folded conformation of the flagellin protein and contribute to the mechanical stability and propulsive force of the flagella fibers.     

Identification of glycosylation genes and glycosylated amino acids of flagellin in Pseudomonas syringae pv. tabaci

A glycosylation island is a genetic region required for glycosylation. The glycosylation island of flagellin in Pseudomonas syringae pv. tabaci 6605 consists of three orfs: orf1, orf2 and orf3. Orf1 and orf2 encode putative glycosyltransferases, and their deletion mutants, Deltaorf1 and Deltaorf2, exhibit deficient flagellin glycosylation or produce partially glycosylated flagellin respectively. Digestion of glycosylated flagellin from wild-type bacteria and non-glycosylated flagellin from Deltaorf1 mutant using aspartic N-peptidase and subsequent HPLC analysis revealed candidate glycosylated amino acids. By generation of site-directed Ser/Ala-substituted mutants, all glycosylated amino acid residues were identified at positions 143, 164, 176, 183, 193 and 201. Matrix-assisted laser desorption/ionization time of flight (MALDI-TOF) mass spectrometry (MS) analysis revealed that each glycan was about 540 Da. While all glycosylation-defective mutants retained swimming ability, swarming ability was reduced in the Deltaorf1, Deltaorf2 and Ser/Ala-substituted mutants. All glycosylation mutants were also found to be impaired in the ability to adhere to a polystyrene surface and in the ability to cause disease in tobacco. Based on the predicted tertiary structure of flagellin, S176 and S183 are expected to be located on most external surface of the flagellum. Thus the effect of Ala-substitution of these serines is stronger than that of other serines. These results suggest that glycosylation of flagellin in P. syringae pv. tabaci 6605 is required for bacterial virulence. It is also possible that glycosylation of flagellin may mask elicitor function of flagellin molecule.     

Role of flagellar hydrogen bonding in Salmonella motility and flagellar polymorphic transition

Bacterial flagellar filaments are assembled by tens of thousands flagellin subunits, forming 11 helically arranged protofilaments. Each protofilament can take either of the two bistable forms L‐type or R‐type, having slightly different conformations and inter‐protofilaments interactions. By mixing different ratios of L‐type and R‐type protofilaments, flagella adopt multiple filament polymorphs and promote bacterial motility. In this study, we investigated the hydrogen bonding networks at the flagellin crystal packing interface in Salmonella enterica serovar typhimurium (S. typhimurium) by site‐directed mutagenesis of each hydrogen bonded residue. We identified three flagellin mutants D108A, N133A and D152A that were non‐motile despite their fully assembled flagella. Mutants D108A and D152A trapped their flagellar filament into inflexible right‐handed polymorphs, which resemble the previously predicted 3L/8R and 4L/7R helical forms in Calladine’s model but have never been reported in vivo. Mutant N133A produces floppy flagella that transform flagellar polymorphs in a disordered manner, preventing the formation of flagellar bundles. Further, we found that the hydrogen bonding interactions around these residues are conserved and coupled to flagellin L/R transition. Therefore, we demonstrate that the hydrogen bonding networks formed around flagellin residues D108, N133 and D152 greatly contribute to flagellar bending, flexibility, polymorphisms and bacterial motility.     

Terminal regions of flagellin are disordered in solution

Limited proteolysis of flagellin from Salmonella typhimurium SJW1103 by subtilisin, trypsin and thermolysin results in homologous degradation patterns. The terminal regions of flagellin are very sensitive to proteolysis. These parts are degraded into small oligopeptides at the very early stage of a mild digestion that yields a relatively stable fragment with a molecular weight of 40,000. Further proteolytic degradation results in a stable 27,000 Mr fragment. The 40,000 Mr tryptic fragment has been identified as residues 67 to 446 of the flagellin sequence, while the 27,000 Mr fragment involves the 179 to 418 segment. The NH2-terminal sequence positions for the corresponding fragments produced by subtilisin are 60 and 174 for the 40,000 Mr and 27,000 Mr fragments, respectively. The fragments lost their polymerizing ability. Structural properties of flagellin and its 40,000 Mr tryptic fragment were compared by circular dichroism spectroscopy and differential scanning calorimetry. Analysis of the calorimetric melting profiles suggests that terminal parts of flagellin have no significant internal stability and they are in extensive contact with water. However, these regions contain some secondary structure, probably alpha-helices, as revealed by comparison of the circular dichroic spectra in the far-ultraviolet region. Our results indicate that, although the terminal regions of flagellin may contain some alpha-helical secondary structure of marginal stability, they have no compact ordered tertiary structure in solution. On the contrary, the central region of the molecule involves at least two compact structural units.